CircleBack

Inspiration## Inspiration

Transitioning into Cambridge A Levels from a CBSE background, I faced a steep hurdle: entering a highly competitive environment with zero prior programming experience in Computer Science or Python. Compounding this, my school was managing its very first Cambridge AS/A Level batch—meaning there were no seniors for guidance, and I found myself as the sole Computer Science student in my entire class.

With only 4 months to prepare for the rigorous exam format, I quickly realized that while practicing past papers is crucial, obtaining personalized, targeted feedback on exact examiner keywords is the real game-changer. Teachers cannot realistically mark dozens of home-practice papers for every student[cite: 1]. This personal challenge inspired me to develop CircleBack—an automated platform dedicated to returning to challenging problems until they become absolute strengths[cite: 1].

What it does

CircleBack is an autonomous, AI-driven evaluation agent built to support intelligent, independent exam preparation[cite: 1]. Built around the 80/20 learning principle, the platform assists students by focusing their energy strictly on converting their weakest analytical areas into high-scoring proficiencies[cite: 1].

Students seamlessly navigate the platform to pick their examination board, subject, specific past paper, and variant[cite: 1]. Once they attempt a structured question, the backend AI evaluation agent cross-references the student's raw text response directly against official mark schemes[cite: 1]. Instead of general feedback, the agent delivers an exact score break down (e.g., $1/3$ marks), explicitly reveals missing keywords (such as catching when a student writes "dilute water" instead of the required "distilled water"), and surfaces analytical trends on their progress dashboard to track study streaks[cite: 1].

How we built it

The core operational prototype was engineered inside the Replit platform, leveraging a Node.js runtime framework integrated with advanced language models to perform context-aware text grading[cite: 1]. To accommodate the scale needed for multi-board curriculum alignment, the system maps out structured relational hierarchies (Board → Subject → Paper → Variant) ensuring students can query exact topical distributions easily[cite: 1].

Challenges we faced

As a beginner programmer who had to rapidly pick up system development tools, managing asynchronous data flows and context validation between student inputs and rigid marking schemes was highly challenging[cite: 1]. Building out the complex, gamified profile tracking system (inspired by motivating platforms like Duolingo) required precise data management to display live progress analytics, streak calculations, and performance metrics accurately[cite: 1].

Accomplishments that we're proud of

We successfully transformed an ambitious educational idea into a fully functional, live-hosted web prototype that has already been tested by over 100 active users[cite: 1]. Proving that structured, AI-assisted evaluation can dramatically reduce manual teacher workload while keeping independent students motivated is a massive milestone for us[cite: 1].

What we learned

Through developing this agent architecture, I mastered fundamental systems engineering concepts, real-world API implementation, and iterative prompt management[cite: 1]. It taught me how to leverage intelligent systems to solve deeply rooted structural problems—a foundational engineering mindset that directly translates to my long-term goals in robotics and autonomous system design[cite: 1].

What's next for CircleBack: Autonomous Exam Evaluation Agent

Our long-term vision is to transition CircleBack from a standalone prototype into a universally trusted, board-aligned ecosystem[cite: 1]. We aim to forge official partnerships with major international examination boards (like Cambridge and CBSE) to securely train our specialized evaluation agents directly on official, proprietary marking schemes[cite: 1]. By engineering enterprise-level data scaling, we aim to deliver trustworthy, highly accurate academic evaluation to a global student audience, helping fulfill the promise of SDG 4: Quality Education[cite: 1].

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for CircleBack

Built With

• chatgpt
• node.js
• replit